9.1 What is DiSTATIS?

Purpose:
(DISTATIS) 3-way multidimensional scaling is a statistical technique that takes root in distance analysis. DISTATIS can be used to analyze both observations and variables where we can see the difference between observations and variables per the a-priori design (pre-assign group). DISTATIS is very similar to clustering methods such as K-Means or Hclust and we can indeed use these methods along side to evaluate DISTATIS results. The beauty of DISTATIS lies in its ability to project The Compromise where variables, observation and their scaled distance are all in the same space. Here we will be using Euclidean Distance

Author’s Notes:
DISTATIS is a very generalized method that shows the hidden complexity and versatility of Multidimensional Scaling. We all know that distances (Euclidean, Mahalanobis, Hellinger, etc.) are on of the cornerstone of statistics. DISTATIS explore the relationships in data using their relative distance between observations and variables. DISTATIS and its siblings methods can be found in brain imaging studies, sorting tasks, classification tasks, sentiment analysis. We can use random effect models to predict or project supplementary data points (novel/test data set).